INDEX

Explanations

words related to places or organizations with "AP" in them

oai_token-act-pair · gpt-3.5-turbo

technical jargon or terms specific to a particular domain, especially related to the arts and media

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

abase

-0.65

 Woods

-0.63

quez

-0.61

rador

-0.59

 Blumenthal

-0.59

 ecology

-0.58

marg

-0.58

 Maduro

-0.58

pole

-0.56

cel

-0.56

POSITIVE LOGITS

haps

0.74

 COURT

0.72

anwhile

0.72

ESE

0.71

urai

0.71

Ã£

0.69

earances

0.69

ople

0.69

monary

0.69

emale

0.68

Activations Density 0.106%

words related to places or organizations with "AP" in them

technical jargon or terms specific to a particular domain, especially related to the arts and media

No Comments

No Known Activations